Variational approach to vibrations of frames with inclined members

Variational approach to vibrations of frames with inclined members

Applied Acoustics 74 (2013) 325–334 Contents lists available at SciVerse ScienceDirect Applied Acoustics journal homepage: www.elsevier.com/locate/a...

627KB Sizes 0 Downloads 64 Views

Applied Acoustics 74 (2013) 325–334

Contents lists available at SciVerse ScienceDirect

Applied Acoustics journal homepage: www.elsevier.com/locate/apacoust

Technical Note

Variational approach to vibrations of frames with inclined members Ricardo Oscar Grossi ⇑, Carlos Marcelo Albarracín Facultad de Ingeniería, Universidad Nacional de Salta, Avenida Bolivia 5150, 4400 Salta, Argentina

a r t i c l e

i n f o

Article history: Received 23 March 2012 Received in revised form 29 June 2012 Accepted 17 July 2012 Available online 24 October 2012 Keywords: Vibrations Frames Inclined members Elastic restrains Functionals Hamilton’s principle

a b s t r a c t This paper deals with the use of calculus of variations to derive the boundary value problems which describe the dynamical behaviour of two and three-bar frames with inclined members, and with ends and intermediate points elastically restrained. The determination of exact eigenfrequencies and modes in the case of free vibrations is included. A rigorous and complete development is presented. First, a brief description of textbooks and papers previously published is included. Second the variational formulation of the problem is presented. Third, the Hamilton principle is rigorously stated and the corresponding boundary value problem is obtained. Finally, the method of separation of variables is used for the determination of the exact frequencies and mode shapes. In order to obtain an indication of the accuracy of the developed mathematical model, some cases available in the literature have been considered and a special code implementing the finite element method have been developed and used. New results are presented for different geometrical configurations and mechanical parameters. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction There are two basic approaches to obtaining the boundary problems that describe the dynamical behaviour of mechanical systems. One is based on Newton’s law of motion; the other is to use Hamilton’s principle to determine the trajectory of the system which minimises the action integral. The first approach is more intuitive but it is not adequate for systems where there are many degrees of freedoms, interacting members and deformable solids. In structural dynamics, the variational approach is extremely important and Hamilton’s principle provides a straightforward method for determining the boundary value problems that describe the dynamical behaviour of mechanical systems. The use of the mentioned principle and the techniques of the calculus of variations constitute an excellent procedure in the case of the determination of the equations of motion, the boundary conditions and the intermediate conditions of systems with many degrees of freedom and many interacting members. Thus, this approach is particularly useful for frame structures with inclined members, whose ends and intermediate points, are elastically restrained against rotation and translation. Substantial literature has been devoted to the theory and applications of the calculus of variations. For instance, the textbooks [1– 8] present the theoretical aspects of the mentioned discipline and some of these works include applications in the statics and dynamics of structural elements such as beams and plates.

⇑ Corresponding author. Fax: +54 0387 4255351. E-mail address: [email protected] (R.O. Grossi). 0003-682X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apacoust.2012.07.014

On the other hand several textbooks and monographs deal with vibrating frames [9–15]. Also, there has been extensive research into the vibration of frame structures and many different configurations and complexities, have been treated. The majority of this work has been in the area of closed frames [16–25]. A number of papers have been published on vibrating planar frame structures with elastically restrained ends. Filipich and Laura [26,27] analysed the in-plane vibrations of portal frames with elastically restrained ends. Laura, Valerga and Filipich [28] dealt with the determination of the fundamental frequency of a frame elastically restrained at the ends, carrying concentrated masses. Albarracín and Grossi [29] dealt with the exact determination of eigenfrequencies of a frame which consists of a beam supported by a column, with intermediate elastic constraints and ends elastically restrained. Grossi and Albarracín [30] used the calculus of variations to derive the boundary problems that describe the dynamical behaviour of portal frames, with ends and intermediate points elastically restrained and determined the exact frequencies and mode shapes. There is only a limited amount of information on the vibration of elastically restrained frames with inclined members. The book by Chin Hao Chang [15] includes the mechanics of elastic structures with inclined members. A model established for the vibration of inclined bars is applied to frames with inclined members. The cases of all joints hinged, two ends fixed and the central joint hinged, and all joints rigidly connected, are investigated. Nevertheless, the case of ends and intermediate points elastically restrained against rotation and translation has not been treated. The aim of the present paper is to investigate the natural frequencies and mode shapes of two and three-bar frames with

326

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

Nomenclature Ai D(F) Da(F) EiIi F li ðiÞ rj R ðiÞ tj t TS US ui ~i u wi ~i w xi

X(G), Y(G) u ~ u v v~ w ~ w

cross-sectional area of the ith beam space of admissible functions of functional F space of admissible directions flexural rigidity of the ith beam Energy functional length of the ith beam spring constant of the jth rotational restraint set of real numbers spring constant of the jth translational restraint time kinetic energy of the mechanical system strain energy of the mechanical system axial displacement along the ith beam admissible direction at ui lateral deflection along the ith beam admissible direction at wi coordinate along the ith beam axis

global coordinates vector (u1, u2, u3) ~2 ; u ~3 Þ ~1 ; u vector ðu vector (u, w) ~ ; wÞ ~ vector ðu vector (w1, w2, w3) ~ 2; w ~ 3Þ ~ 1; w vector ðw ai angles of inclination ~ variation of functional F dFðw; wÞ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi k non-dimensional frequency parameter 4 ðq0 A0 =E0 I0 Þx2i l0

x Xi

qi U W UW

inclined members with ends and intermediate points elastically restrained against rotation and translation, intending the developments within each section to be rigorous and complete. A condensed notation has been implemented by introducing several differential operators. The use of the proposed condensed notation avoids complicated and obscure formulae and allows including all the analytical details. The vague ‘‘operator’’ d which leads to the extensively used mechanical ‘‘d method’’, constitutes a sometimes convenient shorthand tactic, but its lack of rigour can be a source of confusion and can arise as a disadvantage. Consequently, one of the motivations of the present paper is to present a rigorous procedure, by introducing the functional adequate to the formulation of the Hamilton’s principle and the corresponding spaces of admissible functions and admissible directions. Hamilton’s principle requires that between times ta and tb, at which the positions of the mechanical system are known, it should execute a motion which makes stationary the functional given by the Hamilton’s action integral

FðwÞ ¼

Z

tb

ðT S ðwÞ  U S ðwÞÞdt;

ð1Þ

ta

where L = TS  US is the well known Lagrangian for the motion. It must be noted that the frame is a structural system which cannot be studied using as admissible functions neither w(, t) 2 C4[0, l] (used in beams) nor wð; ; tÞ 2 C 4 ðXÞ; X # R2 (used in plates). So, it is necessary to clearly determine the space of admissible functions D of the functional defined by (1) and the space Da of admissible directions. The procedure adopted is particularly important in the determination of the analytical expressions of the corresponding natural boundary conditions and the intermediate conditions. In the present paper the method of separation of variables is used for the determination of the exact frequencies and mode shapes. Tables are given for frequencies and two-dimensional plots are given for mode shapes in some selected cases. In order to obtain an indication of the accuracy of the developed mathematical model, some particular cases available in the literature has been considered, and a particular code with the application of the finite element method has been developed. The present paper is organised in the following way. First the brief story stated above. In Section 2, the variational formulation of the problem is described. In Section 3 the Hamilton principle is rigorously stated by introducing adequate vectorial functions and a particular product space. Also the corresponding boundary

circular frequency of the structure (rad/unit time) [0, li]  [ta, tb] material density of the ith beam C2(X1)  C2(X2)  C2(X3) C4(X1)  C4(X2)  C4(X3) linear product space

value problem is obtained. In Section 4 the implementation of the method of separation of variables is described for the determination of the results of exact frequencies and mode shapes. The paper is concluded in Section 5 with some discussions. 2. Variational formulation of the problem The frame considered is composed by three inclined beams of lengths l1, l2 and l3 respectively as it is shown in Fig. 1, where the coordinate systems has been adopted to help the analytical developments. The angles of inclination are given by ai, i = 1, 2, 3. The behaviours of the individual members of the frame are assumed to be governed by Euler’s beam theory with the axial deformations effects included. In practice, the frame corners and ends may experience partial resistance to rotation and translation, and this situation can be modelled by considering ends and joints elastically restrained against rotation and translation. For this reason, the following elastic restraints have been included: (1) Four rotational restraints characterised by the spring constants ri, i = 1, . . ., 4. (2) Eight translational restraints characterised by the spring constants tix, tiy, i = 1, 2, 3, 4. The longitudinal and transverse displacements of the beams are related to their local axes at any time t and described respectively by the functions

ui ðxi ; tÞ; wi ðxi ; tÞ;

xi 2 ½0; li ;

i ¼ 1; 2; 3:

ð2Þ

In the joints of connection, there exist constrained conditions which lead to the following compatibility equations (see Fig. 1):

cos a1 u1 ðl1 ; tÞ  sin a1 w1 ðl1 ; tÞ ¼ cos a2 u2 ð0; tÞ  sin a2 w2 ð0; tÞ; ð3Þ sin a1 u1 ðl1 ; tÞ þ cos a1 w1 ðl1 ; tÞ ¼ sin a2 u2 ð0; tÞ þ cos a2 w2 ð0; tÞ; ð4Þ @w1 @w2 ðl1 ; tÞ ¼ ð0; tÞ; @x1 @x2

ð5Þ

cos a2 u2 ðl2 ; tÞ  sin a2 w2 ðl2 ; tÞ ¼ cos a3 u3 ð0; tÞ  sin a3 w3 ð0; tÞ; ð6Þ sin a2 u2 ðl2 ; tÞ þ cos a2 w2 ðl2 ; tÞ ¼ sin a3 u3 ð0; tÞ þ cos a3 w3 ð0; tÞ; ð7Þ @w2 @w3 ðl2 ; tÞ ¼ ð0; tÞ: @x2 @x3

ð8Þ

At time t, the kinetic energy of the mechanical system under study is given by

327

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

l2

Y (G )

w2 x1 , u1 r2

x2 , u2 , w3

t3 y

E2 , I 2 , ρ 2 , A2 t2 y

t3 x

α2

t2 x

E3 , I 3 , ρ3 , A3

E1 , I1 , ρ1 , A1

l1

α3

r3

l3

α1 w1

t1x

r4

r1

t4 x t4 y

t1 y

x3 , u3

X (G )

Fig. 1. Frame with inclined members and with ends and intermediate points elastically restrained.

TS ¼

Z

3 1X 2 i¼1

li

qi Ai

0

 2  2 ! @ui @wi dxi ; ðxi ; tÞ þ ðxi ; tÞ @t @t

ð9Þ

where qiAi denotes the mass per unit length of the ith member of the frame. The total potential energy due to the elastic deformation of the beams, the springs at the ends restraints and the springs at the intermediate restraints is given by: 3 1X US ¼ 2 i¼1

Z

li

0

!2 1  2 2 @u @ w i i @Ei Ai ðxi ; tÞ þ Ei Ii ðxi ; tÞ Adxi @xi @x2i

tb

ta

2

Z 0

li

qi Ai

"

@ui ðxi ;tÞ @t

2

 

Z

1 2 Z

li

tb

ð10Þ

 2 # 3 Z tb @wi 1X dxi dt  ðxi ;tÞ 2 i¼1 ta @t !2 3 2 @ wi ðxi ;tÞ 5dxi dt @x2i þ

ta

" ð4Þ

ð4Þ

ð4Þ

k11 u23 ðl3 ;tÞ þ k22 w23 ðl3 ;tÞ þ k33



w ¼ ðw1 ; w2 ; w3 Þ:

ð12Þ

consists of all the n-tuples u = (u1, u2, . . . , un) where

 2 4Ei Ai @ui ðxi ;tÞ þ Ei Ii @x i 0 " #  2 3 Z tb X 1 ðiÞ ðiÞ ðiÞ @wi ðiÞ  k11 u2i ð0;tÞ þ k22 w2i ð0;tÞ þ k33 ð0;tÞ þ 2k12 ui ð0;tÞwi ð0;tÞ dt 2 i¼1 ta @xi



u ¼ ðu1 ; u2 ; u3 Þ;

V ¼ V1  V2      Vn

where EiIi denotes the flexural rigidity of the ith member of the ðlÞ frame and the rigidities kij generated by the elastic restraints are defined in Appendix A. So, the energy functional to be considered is given by Z

i ¼ 1; 2; 3;

Definition. Let V1, V2, . . ., Vn, be linear spaces. The product space

ðiÞ

3 1X 2 i¼1

wi ;

Now we recall the definition of product vector spaces, see for instance [31,32].

þ 2k12 ui ð0; tÞwi ð0; tÞ "  2 1 ð4Þ 2 ð4Þ ð4Þ @w3 þ k11 u3 ðl3 ; tÞ þ k22 w23 ðl3 ; tÞ þ k33 ðl3 ; tÞ 2 @x3



ui ;

v ¼ ðu; wÞ;

"  2 3 1X ðiÞ ðiÞ ðiÞ @wi þ k11 u2i ð0; tÞ þ k22 w2i ð0; tÞ þ k33 ð0; tÞ 2 i¼1 @xi

ð4Þ

A simple examination of (11) reveals that it defines a functional F which depends on the six real functions

so, it is convenient to introduce the following vectorial functions:

0

þ 2k12 u3 ðl3 ; tÞw3 ðl3 ; tÞ;

3. Product spaces and the concept of variation of the energy functional

# 2 @w3 ð4Þ ðl3 ;tÞ þ 2k12 u3 ðl3 ;tÞw3 ðl3 ;tÞ dt: @x3

ð11Þ

ui 2 V i

for i ¼ 1; 2; . . . ; n:

The algebraic operations:

u þ v ¼ ðu1 þ v 1 ; u2 þ v 2 ; . . . ; un þ v n Þ; au ¼ ðau1 ; au2 ; . . . ; aun Þ;

8a 2 R;

8u; v 2 V; 8u 2 V;

transforms the space V in a linear space. It is worth pointing out that the notion of product space allows to clearly specify the degree of smoothness of the functions involved in (11). For this purpose, we will make the following assumptions:

ui ðxi ; tÞ 2 C 2 ðXi Þ and wi ðxi ; tÞ 2 C 4 ðXi Þ where Xi ¼ ½0; li   ½ta ; t b ; i ¼ 1; 2; 3: Now if we introduce the spaces

U ¼ C 2 ðX1 Þ  C 2 ðX2 Þ  C 2 ðX3 Þ and

ð13Þ

328

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

W ¼ C 4 ðX1 Þ  C 4 ðX2 Þ  C 4 ðX3 Þ;

ð14Þ

it is immediately that u 2 U, w 2 W and v 2 U  W. Consequently, the functional defined by (11) depends on v and it can be written F = F(v). It is quite easy to check that the product space U  W can be transformed in a linear space with the following operations:

ðuð1Þ ; wð1Þ Þ þ ðuð2Þ ; wð2Þ Þ ¼ ðuð1Þ þ uð2Þ ; wð1Þ þ wð2Þ Þ;

~i ðai ; tÞ ¼ ~ i ðai ; tÞ ¼ u w

3 Z X

Z



tb

li

"

qi Ai

0

ð15Þ

Since the domain of the functional F = F(v) has been determined, it is possible to define the variation of the functional by

  dF ~ Þ ; ðv þ ev de e¼0

ð16Þ

~ Þ denotes the variation of functional F in the point where dFðv ; v ~ 2 Da ðFÞ. The admissible directions v 2 D(F) and in the direction v v~ at v 2 D(F) are those for which v þ ev~ 2 DðFÞ; 8 sufficiently small e and exists dFðv ; v~ Þ. In consequence, in view of (15), v~ is an admis~ 2 Da ðFÞ where sible direction at v for D(F) if and only if v

~; v ~ 2 U  W; fulfils the compatibility conditions Da ðFÞ ¼ fv ~ ðxi ; t b Þ ¼ 0; 8xi 2 ½0; li ; i ¼ 1; 2; 3g: ð17Þ ~ ðxi ; t a Þ ¼ v ð3Þ—ð8Þ and v The application of (16) to the functional defined by (11) leads to a rather lengthy algebraic procedure which can be condensed by introducing the following differential operators:

Ei Ai

@ 2 ui @ 2 ui ðxi ; tÞ  qi Ai 2 ðxi ; tÞ ¼ 0; 2 @xi @t

Ei Ii

@ 4 wi @ 2 wi ðx ; tÞ þ q A ðxi ; tÞ ¼ 0; i i i @x4i @t 2

~Þ ¼ dFðv ; v

4 Z X

þ

tb

ta



li

!

q i Ai

0

þ

i¼1

þ

i¼1

tb

ta

4 Z X

~ i ðai ; tÞQ i ðai ; tÞdt u

ta

4 Z X

tb

~ i ðai ; tÞRi ðai ; tÞdt: w

!

@ 2 ui @ 2 ui ~i þ  E i Ai 2 u @xi @t2

4 X ~i @w ðai ; tÞP i ðai ; tÞdt þ @xi i¼1

qi Ai Z

ð27Þ

ta

X ¼ X i ¼ xi =li ; Ri ¼

r i li ; Ei I i

i ¼ 1; . . . ; 3;

i ¼ 1; . . . ; 4;

3

#

@ 2 wi @ 4 wi ~ i dxi dt þ Ei I i 4 w @xi @t2

tb

t ix li ; Ei Ii

T ix ¼

t ix li1 ; Ei1 Ii1

~ i ðai ; tÞQ i ðai ; tÞdt u

ta

li Lij ¼ ; lj Eij ¼

tb

~ i ðai ; tÞRi ðai ; tÞdt; w

ta

ð21Þ

J0 ¼

Ei ; Ej I0 2

A0 l 0

;

ð28aÞ

l4 ¼ l3 ;

E4 I4 ¼ E3 I3 ;

T iy ¼

t iy li ; Ei I i

i ¼ 1; 2;

ð28cÞ

3

T iy ¼

t iy li1 ; Ei1 Ii1

i ¼ 3; 4;

Ai ; Aj

ði; jÞ 2 fð1; 0Þ; ð2; 0Þ; ð3; 0Þ; ð2; 1Þ; ð3; 2Þg; ð28eÞ

Iij ¼

Ii ; Ij

ði; jÞ 2 fð1; 0Þ; ð2; 0Þ; ð3; 0Þ; ð1; 2Þ; ð2; 3Þg; ð28fÞ

Ji ¼

Ii 2

Ai l i

;

qi0 ¼

qi ; i ¼ 1; 2; 3; q0

where l0, q0, A0, E0 and I0 are generic parameters. By assuming separable solutions in the form

~ Þ ¼ 0; dFðv ; v

ui ðX; tÞ ¼ U i ðXÞ cos xt;

ð22Þ

~ 2 Da ðFÞ which verify the conditions Restricting attention to those v

ð28dÞ

Aij ¼

~4 ¼ w ~ 3; u ~4 ¼ u ~3 . where ai ¼ 0; i ¼ 1; 2; 3; a4 ¼ l3 ; w Now the stationary condition which corresponds to Hamilton’s principle takes the following rigorous form:

8v~ 2 Da ðFÞ:

ð28bÞ

3

T ix ¼

3

ta

4 Z X

tb

ð20Þ

tb

"

ð26Þ

~ i ðai ; tÞ; u ~ i ðai ; tÞ; @@xw~ i ðai ; tÞ; 8t 2 ðt a ; t b Þ; i ¼ 1; Since the functions w i . . . ; 4, are smooth and arbitrary, the stationary condition (22) leads to the corresponding natural boundary conditions. To reduce the number of defining variables and parameters the following non-dimensional variables and parameters are introduced:

where j = i if i = 1, 2, 3 and j = i  1 if i = 4. Thus we have (see Appendix A for more details):

Z

i

ð19Þ

@ 3 wj @ 3 wj1 ðiÞ ðiÞ ðx; tÞ  k22 wj ðx; tÞ  k12 uj ðx; tÞ þ g i 3 @xj @x3j1

i¼1

8xi 2 ð0; li Þ;

ð25Þ

~i @w ðai ; tÞPi ðai ; tÞdt @xi

4 Z X

i¼1

@uj1 ðx þ lj1 ; tÞ; @xj1

8t 2 ½t a ; tb ;

Now it is possible to remove the restrictions (23), and since the functions ui must satisfy Eq. (25) and the functions wi Eq. (26), the expression (21) reduces to

þ

ðiÞ

i

8t 2 ½ta ; t b :

¼ 1; 2; 3;

ð18Þ

þ lj1 ; tÞ  k11 uj ðx; tÞ  k12 wj ðx; tÞ;

8xi 2 ð0; li Þ;

¼ 1; 2; 3;

@uj @ 3 wj1 @uj1 ðx; tÞ þ ei ðx þ lj1 ; tÞ þ fi ðx @xj @xj1 @x3j1

3 Z X ~Þ ¼  dFðv ; v

! # @ 2 wi @ 4 wi ~ i dxi dt þ Ei I i w 2 4 @xi @t

~ i and w ~ i are arbitrary smooth functions and verify the condiSince u tions (23), the fundamental lemma of the calculus of variations can be applied to Eq. (24) to conclude that the functions ui and wi must respectively satisfy the following differential equations:

i¼1

 ðx þ lj1 ; tÞ þ hi

q i Ai

~ 2 Da ðFÞ: ¼ 0; 8v

i¼1

@ 2 wj @ 2 wj1 ðiÞ @wj Pi ðx; tÞ ¼ bi ðx; tÞ þ ci ðx þ lj1 ; tÞ  k33 ðx; tÞ; 2 @xj @xj @x2j1

Ri ðx; tÞ ¼ bi

! @ 2 ui @ 2 ui ~i þ  E i Ai 2 u 2 @xi @t

ð24Þ

DðFÞ ¼ fv ; v 2 U  W; fulfils the compatility conditions ð3Þ—ð8Þ and has prescribed values at t ¼ ta and t ¼ t b g:

ðiÞ

ð23Þ

ta

i¼1

where u(1), u(2) 2 U, w(1), w(2) 2 W and c 2 R. Now it is possible to clearly specify the domain of admissible functions which is given by

Q i ðx; tÞ ¼ di

i

we have from (21) and (22) that

¼ ðcu; cwÞ;

~Þ ¼ dFðv ; v

8t 2 ðt a ; tb Þ;

¼ 1; . . . ; 4;

~Þ ¼ dFðv ; v

cðu; wÞ

~i @w ðai ; tÞ ¼ 0; @xi

wi ðX; tÞ ¼ W i ðXÞ cos xt;

ð28gÞ

ð29Þ i ¼ 1; 2; 3;

X 2 ½0; 1;

t P 0;

ð30Þ

329

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

the natural boundary conditions and the compatibility conditions can be written as: 2

d W1

ð1Þ dW 1 ð0Þ  K 33 ð0Þ ¼ 0; dX dX 2 dU 1 ð1Þ ð1Þ ð0Þ  J 1 K 11 U 1 ð0Þ  J1 K 12 W 1 ð0Þ ¼ 0; dX 3 d W1 ð1Þ ð1Þ ð0Þ þ K 22 W 1 ð0Þ þ K 12 U 1 ð0Þ ¼ 0; dX 3 2 d W 3 ð1Þ ð4Þ dW 3 ð1Þ þ K 33 ¼ 0; dX 3 dX 2 dU 3 ð4Þ ð4Þ ð1Þ þ J 3 K 11 U 3 ð1Þ þ J 3 K 12 W 3 ð1Þ ¼ 0; dX 3 d W3 ð4Þ ð4Þ ð1Þ  K 22 W 3 ð1Þ  K 12 U 3 ð1Þ ¼ 0; dX 3 cos a1 U 1 ð1Þ  sin a1 W 1 ð1Þ  cos a2 U 2 ð0Þ þ sin a2 W 2 ð0Þ ¼ 0; sin a1 U 1 ð1Þ þ cos a1 W 1 ð1Þ  sin a2 U 2 ð0Þ  cos a2 W 2 ð0Þ ¼ 0; dW 1 dW 2 L21 ð1Þ  ð0Þ ¼ 0; dX dX cos a2 U 2 ð1Þ  sin a2 W 2 ð1Þ  cos a3 U 3 ð0Þ þ sin a3 W 3 ð0Þ ¼ 0; sin a2 U 2 ð1Þ þ cos a2 W 2 ð1Þ  sin a3 U 3 ð0Þ  cos a3 W 3 ð0Þ ¼ 0; dW 2 dW 3 L32 ð1Þ  ð0Þ ¼ 0; dX dX 2 2 d W2 d W1 ð2Þ dW 2 ð0Þ  K 33 ð1Þ ¼ 0; ð0Þ  E12 I12 L221 2 dX dX dX 2 2 2 d W3 d W2 ð3Þ dW 3 ð0Þ  K 33 ð1Þ ¼ 0; ð0Þ  E23 I23 L232 2 dX dX dX 2 ð2Þ

ð31Þ ð32Þ

3

 a1 Þ

dX 3

dX ¼ 0;

3

ð1Þ 

Using the well-known separation of variables method, the solutions of Eqs. (49) and (50) are respectively assumed to be of the form:

ð37Þ ð38Þ ð39Þ ð40Þ ð41Þ ð42Þ

dX 3

ð44Þ

ð46Þ ð2Þ

ð0Þ þ K 22 W 2 ð0Þ þ K 12 U 2 ð0Þ  cosða2  a1 ÞE12 I12 L321 3



d W1

dX 3 ¼ 0;

ð1Þ þ G3

ð51Þ

 pffiffiffiffiffiffi   pffiffiffiffiffiffi  ðiÞ ðiÞ ðiÞ W i ðXÞ ¼ c1 cosh k 4 C 1i X þ c2 sinh k 4 C 1i X þ c3  pffiffiffiffiffiffi   pffiffiffiffiffiffi  ðiÞ  cos k 4 C 1i X þ c4 sin k 4 C 1i X ; i ¼ 1; 2; 3:

ð52Þ

Replacing Eqs. (51) and (52) into the conditions (31)–(48), we obðiÞ tain a set of eighteen homogeneous equations in the constants cj ðiÞ and dj . Since the system is homogeneous, for existence of a nontrivial solution, the determinant of coefficients must be equal to zero. This procedure yields the frequency equation:

GðR; Tx ; Ty ; kÞ ¼ 0;

G2 dU 2 ð1Þ J 2 dX

ð2Þ

 pffiffiffiffiffiffi   pffiffiffiffiffiffi  ðiÞ ðiÞ U i ðXÞ ¼ d1 cos k2 C 2i X þ d2 sin k2 C 2i X ;

ð43Þ

3

d W2

x2 l40 :

ð36Þ

G1 dU 1 ð1Þ  ð1Þ J 1 dX

3

E0 I0

4. Numerical results

ð45Þ

d W2

q0 A0

and k4

ð35Þ

A32 dU 3 ð3Þ ð3Þ ð0Þ  K 11 U 3 ð0Þ  K 12 W 3 ð0Þ  sinða2  a3 Þ L32 E23 J2 dX 

¼

C 2i ¼ ðqi0 Ai0 ÞðEi0 Ai0 Þ1 J 0 L2i0

ð34Þ

ð33Þ

ð2Þ

¼ 0;

C 1i ¼ ðqi0 Ai0 ÞðEi0 Ii0 Þ1 L4i0 ;

In conclusion, the boundary problem which describes the natural vibrations of the mechanical system under study is given by Eqs. (49) and (50) and the boundary or compatibility conditions (31)– (48). It must be noted that the angles of inclination must verify the conditions a1 – 0 and a2 – np/2, n = 1, 2, . . . .

K K A21 dU 2 ð0Þ  3 11 U 2 ð0Þ  3 12 W 2 ð0Þ þ sinða2 L21 E12 J1 dX L21 E12 I12 L21 E12 I12 d W1

where

E12 L21 dU 1 ð1Þ A21 J 2 dX

where the components of R, Tx and Ty are given by Eqs. (28b–d). In order to obtain an indication of the accuracy of the developed mathematical model, the classical problems analysed in Refs. [9,29,30] have been solved. In all the cases an excellent agreement has been observed. Since the algorithm developed allows and great number of configurations and mechanical characteristics and the number of cases is prohibitively large, results are presented for only a few cases. All the numerical results have been obtained with the following generic parameters: q0 = q1, E0 = E1, I0 = I1, A0 = A1 and l0 = l1. Table 1a depicts values of coefficients ki, i = 1, 2, . . ., 5 where 4 k4i ¼ qE00AI00 x2i l0 ; for a frame with rigidly clamped ends and one inclined member. The geometric and mechanical parameters values are given by:

E12 ¼ E23 ¼ 1; ð47Þ

ð53Þ

I12 ¼ I23 ¼ 1;

L10 ¼ 1; a1 ¼ p=2;

A21 ¼ A32 ¼ 1;

J 1 ¼ 1=30000;

a3 ¼ p=2

3



d W3 dX 3

ð3Þ

ð3Þ

ð0Þ  L332 E23 I23 K 22 W 3 ð0Þ  L332 E23 I23 K 12 U 3 ð0Þ 3

þ cosða2  a3 ÞE23 I23 L332

d W2 dX

3

ð1Þ  G4

and the remaining values are given in Table 1b except the values of a2, J2, J3, L21, L32 and L20 which vary with L30, and are depicted in Table 1c. It can be observed that the values of ki, i = 1, 2, . . ., 5 decrease when the values of the parameter L30 increase. The first five mode

E23 L32 dU 2 ð1Þ A32 J 3 dX

¼ 0;

ð48Þ ðlÞ K ij

ðlÞ

where the coefficients are obtained from the expression of kij replacing tixand tiy respectively by Tix and Tiy. The coefficients Gi are listed in Appendix A. Now, Eqs. (25) and (26) reduce to 2

d Ui

4

ðXÞ þ k C2i U i ðXÞ ¼ 0; 8X 2 ð0; 1Þ; i ¼ 1; 2; 3; dX 2 4 d Wi ðXÞ  k4 C1i W i ðXÞ ¼ 0; 8X 2 ð0; 1Þ; i ¼ 1; 2; 3; dX 4

ð49Þ ð50Þ

Table 1a First five exact values of the frequency parameter k for a frame with one inclined member and ends rigidly clamped, for different values of L30. L30

k1

k2

k3

k4

k5

1 6/5 7/5 8/5 9/5

1.51405646 1.41721109 1.34001905 1.27532429 1.21867577

2.02926280 2.01115388 1.99328729 1.97142612 1.94462321

3.39508597 3.28370976 3.03178225 2.72980005 2.46273729

4.19589016 3.78347996 3.51625304 3.36862073 3.23723455

4.59485951 4.42294976 4.35387959 4.26535142 4.07086141

330

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

shapes which correspond to the case L30 = 9/5 are presented in Fig. 2a–e. Table 2a depicts values of coefficients ki, i = 1, 2, . . ., 5 for a frame with rigidly clamped ends and one inclined member. The geometric and mechanical parameters values are given by:

L21 ¼ 2; E12 ¼ E23 ¼ 1; I12 ¼ I23 ¼ 1; A21 ¼ A32 ¼ 1; J 1 ¼ 1=30000; J 2 ¼ 1=120000; L10 ¼ 1; L20 ¼ 2; a1 ¼ p=2; a2 ¼ 0; and the remaining values are given in Table 2b except the values of J3, L32 and L30 which vary with a3 and are depicted in Table 2c. The first five mode shapes which correspond to the case a3 =  p/4 are presented in Fig. 3a–e. Table 3a depicts values of coefficients ki, i = 1, 2, . . ., 5 for a frame with a fixed geometry and two intermediate points elastically restrained. The geometric and mechanical parameters values are given by:

Table 1c Values of the characteristic parameters a2, J2, J3, L21, L32 and L20 which vary with L30 and correspond to the case depicted in Table 1a. L30

a2

L20 = L21

J2

J3

L32

1 6/5

0.00E+0 7.60E3

1/120000 1/121200

1/30000 1/43200

7/5

7.73E3

1/124800

1/55800

8/5

7.77E3

9/5

7.79E3

2 pffiffiffiffiffiffiffiffiffi 101=5 pffiffiffiffiffiffi 2 26=5 pffiffiffiffiffiffiffiffiffi 109=5 pffiffiffiffiffiffi 2 29=5

1/2 pffiffiffiffiffiffiffiffiffi 6 101=101 pffiffiffiffiffiffi 7 26=52 pffiffiffiffiffiffiffiffiffi 8 109=109 pffiffiffiffiffiffi 9 29=58

1/130800

1/76800

1/139200

1/97200

R1 ¼ R4 ¼ 1; R3 ¼ 0; T 1x ¼ T 4x ¼ T 1y ¼ T 4y ¼ 1; T 2x ¼ T 2y ¼ 0; pffiffiffi pffiffiffi L21 ¼ 2; L32 ¼ 2=2; E12 ¼ E23 ¼ 1; I12 ¼ I23 ¼ 1; A21 ¼ A32 ¼ 1; and the remaining values are given in Table 3b. It can be observed that the values of ki, i = 1, 2, . . ., 5 increase with the increment of the values of the parameters T3x, T3y, and R2. The first five mode shapes which correspond to the case T3x = T3y = R2 = 100, are presented in Fig. 4a–e. Table 4a depicts values of coefficients ki, i = 1, 2, . . ., 5 for a classical portal frame with ends elastically restrained against rotation. The geometric and mechanical parameters values are given by:

R2 ¼ R3 ¼ 0;

L21 ¼ 2;

I12 ¼ I23 ¼ 1;

A21 ¼ A32 ¼ 1;

L32 ¼ 1=2;

L32 ¼ 1;

(b) Second mode shape

(c) Third mode shape

(d) Fourth mode shape

E12 ¼ E23 ¼ 1;

and the remaining values are given in Table 4b. In this case, the values of ki, i = 1, 2, . . ., 5 increase with the increment of the values of the parameters R1 and R4. Comparison of Tables 3a and 4a shows that the translational restraints generally have greater influence on the frequencies than the rotational restraints. The model established for the vibration of a frame with three inclined members can be applied to a two-bar frame. For such a frame it is sufficient to adopt the condition a2 = a3. Table 5a depicts values of coefficients ki, i = 1, 2, . . ., 5 for a twobar frame with ends elastically restrained against rotation and translation. The geometric and mechanical parameters values are given by:

L21 ¼ 1=2;

(a) First mode shape

E12 ¼ E23 ¼ 1;

I12 ¼ I23 ¼ 1;

A21 ¼ A32 ¼ 1;

(e) Fifth mode shape

and the remaining values are given in Table 5b. In this case, the values of ki, i = 1, 2, . . ., 5 increase with the increment of the values of K. The first five mode shapes which correspond to the case K = 1 and K = 1000 are respectively presented in Fig. 5a–e and Fig. 6a–e. For all the cases depicted by Tables 1a–5a, the finite element method was implemented and a special code developed where 25 Euler beam elements have been employed to divide each beam. The numerical values obtained with this method, coincide at least

Table 1b Values of characteristic parameters Ri, Tix, Tiy, Ii0, Ai0, Ei0 and qi0 which correspond to the case depicted in Table 1a. i

Ri

Tix

Tiy

Ii0

Ai0

Ei0

qi0

1 2 3 4

1 0 0 1

1 0 0 1

1 0 0 1

1 1 1 –

1 1 1 –

1 1 1 –

1 1 1 –

Fig. 2. Mode shapes of the frame with the same parameters used in Table 1a in the case L30 = 9/5: (a) first mode shape, (b) second mode shape, (c) third mode shape, (d) fourth mode shape, (e) fifth mode shape.

Table 2a First five exact values of the frequency coefficient k for a frame with rigidly clamped ends and one inclined member, for different values of a3.

a3

k1

k2

k3

k4

k5

p/4 3p/ 8 p/2 5p/ 8 3p/ 4

1.44215592 1.52216252

2.01495355 2.03384784

3.04552404 3.35909676

3.62061372 4.03838813

4.42211965 4.52411483

1.51405646 1.42440808

2.02926280 2.03930622

3.39508597 3.38256465

4.19589016 4.03291179

4.59485951 4.46872486

1.21392548

2.07675290

3.08866802

3.55039980

4.34065004

331

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334 Table 2b Values of characteristic parameters Ri, Tix, Tiy, Ii0, Ai0, Ei0 and qi0 which correspond to the case depicted in Table 2a. i

Ri

Tix

Tiy

Ii0

Ai0

Ei0

qi0

1 2 3 4

1 0 0 1

1 0 0 1

1 0 0 1

1 1 1 –

1 1 1 –

1 1 1 –

1 1 1 –

(a) First mode shape

(c) Third mode shape

Table 2c Values of the characteristic parameters J3, L30 and L32 which vary with a3 and correspond to the case depicted in Table 2a.

a3

J3

L30

L32

p/4 3p/8 p/2 5p/8 3p/4

1.67E5 2.85E5 3.33E5 2.85E5 1.67E5

1.414 1.082 1.000 1.082 1.414

0.707 0.541 0.500 0.541 0.707

(b) Second mode shape

(d) Fourth mode shape

(e) Fifth mode shape Fig. 4. Mode shapes of the frame with the same parameters used in Table 3a in the case T3x = T3y = R2 = 100: (a) first mode shape, (b) second mode shape, (c) third mode shape, (d) fourth mode shape, (e) fifth mode shape.

Table 4a First five exact values of the frequency coefficient k for a classical portal frame with ends elastically restrained against rotation, for different values of the parameters R1 and R4.

(a) First mode shape

(b) Second mode shape

(c) Third mode shape

(d) Fourth mode shape

R1 = R4

k1

k2

k3

k4

k5

0 1 10 100 1000 10000

1.00735667 1.17056652 1.40024036 1.47149835 1.48027594 1.48117443

1.89604323 1.90684805 1.93009678 1.93941103 1.94062957 1.94075519

2.80366323 2.80616330 2.81123764 2.81315615 2.81340241 2.81342774

3.04479119 3.05528110 3.07324651 3.07893735 3.07962756 3.07969804

3.25339525 3.29161526 3.39089853 3.43797790 3.44445612 3.44512809

Table 4b Values of the characteristic parameters Tix, Tiy, Ji, ai, Li0, Ii0, Ai0, Ei0 and qi0 which correspond to the case depicted in Table 4a.

(e) Fifth mode shape Fig. 3. Mode shapes of the frame with the same parameters used in Table 2a in the case a3 = p/4: (a) first mode shape, (b) second mode shape, (c): third mode shape, (d) fourth mode shape, (e) fifth mode shape.

Table 3a First five exact values of the frequency coefficient k for a frame with a fixed geometry and two intermediate points elastically restrained. T3x = T3y = R2 k1 0.1 1.0 10.0 100.0 1000.0

k2

2.05141270 2.05672993 2.10631121 2.46585465 3.01195249

k3

2.74674211 2.76782191 2.87768329 2.99012833 3.69232241

k4

4.15813108 4.17877712 4.27928093 4.32884668 4.35373620

k5

4.40725873 4.41378971 4.49634494 4.75083373 4.86254731

5.56447003 5.56682495 5.58487356 5.63290958 5.73338619

Table 3b Values of the characteristic parameters Ji, ai, Li0, Ii0, Ai0, Ei0 and qi0 which correspond to the case depicted in Table 3a. i

Ji

ai

Li0

Ii0

Ai0

Ei0

qi0

1 2

1/60000 1/120000

p/4

1 1

1 1

1 1

1 1

3

1/60000

p/4

1 pffiffiffi 2 1

1

1

1

1

0

i

Tix

Tiy

Ji

ai

Li0

Ii0

Ai0

Ei0

qi0

1 2 3 4

100 0 0 100

100 0 0 100

1/30000 1/120000 1/30000 –

p/2

1 2 1 –

1 1 1 –

1 1 1 –

1 1 1 –

1 1 1 –

0 p/2 –

Table 5a First five exact values of the frequency coefficient k for a two-beams frame with ends elastically restrained against rotation and translation. The restraint parameter vary with K, where T1x = T1y = T4x = T4y = R1 = R4 = K. K

k1

k2

k3

k4

k5

0 1 10 100 1000 10000

0.00000000 1.72257177 2.42449159 3.45649311 3.87182675 3.91843903

1.87059351 2.45065136 3.14301012 4.01623701 4.60882552 4.70916362

2.80025968 2.90116036 3.55770052 4.57459083 5.92481310 7.01271047

4.07815885 4.33543228 4.76792675 5.32647698 6.78226633 7.72026488

4.42332371 4.61859014 5.06750404 6.06926303 7.64381353 9.19332081

Table 5b Values of the characteristic parameters Tix, Tiy, Ri, Ji, ai, Li0, Ii0, Ai0, Ei0 and qi0 which correspond to the case depicted in Table 5a. i

Tix

Tiy

Ri

Ji

ai

Li0

Ii0

Ai0

Ei0

qi0

1 2 3 4

K 10 0 K

K 100 0 K

K 0 0 K

1/10000 1/2500 1/2500 –

p/4 p/4 p/4

1 1/2 1/2 –

1 1 1 –

1 1 1 –

1 1 1 –

1 1 1 –



332

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

(b) Second mode shape

(a) First mode shape

(d) Fourth mode shape

(c) Third mode shape

(e) Fifth mode shape Fig. 5. Mode shapes of the frame with the same parameters used in Table 5a in the case K = 1: (a) first mode shape, (b) second mode shape, (c) third mode shape, (d) fourth mode shape, (e) fifth mode shape.

up to four decimal digits with the exact values, so they have not been included in the tables.

5. Conclusions In this paper the calculus of variations was used to derive the boundary value problems which describe the dynamical behaviour of two and three-bar frames with inclined members, and with ends and intermediate points elastically restrained against rotation and translation. A rigorous procedure was used by introducing a functional adequate to the formulation of the Hamilton’s principle and the corresponding spaces of functions. The stationary condition for the functional F defined by (11) on the space of admissible functions has been clearly stated by defining the space of admissible functions D(F) and the space Da(F) of admissible directions. It has been shown that the introduction of the vectorial functions (12), the

spaces (13), (14) and the product space U  W, leads to the rigorous definition of the variation of the functional F by means of the derivative (16). A simple, computationally efficient and accurate approach has been developed for the determination of natural frequencies and modal shapes of free vibration of the frames described above. Numerical results for the first five natural frequencies in tabular form and the corresponding mode shapes were included. In order to obtain an indication of the accuracy of the mathematical model obtained, some cases available in the literature have been considered and the finite element method has also been implemented. In all cases excellent agreements have been determined. The algorithm is very general and it is attractive regarding its versatility in handling boundary and intermediate conditions. Besides, it allows taking into account a great variety of complicating effects and the adoption of different generic parameters l0, q0, A0, E0 and I0, leads to different analytical expressions for the restraint parameters and the frequency coefficients.

333

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

(a) First mode shape

(b) Second mode shape

(c) Third mode shape

(d) Fourth mode shape

(e) Fifth mode shape Fig. 6. Mode shapes of the frame with the same parameters used in Table 5a in the case K = 1000: (a) first mode shape, (b) second mode shape, (c) third mode shape, (d) fourth mode shape, (e) fifth mode shape. ðiÞ

Acknowledgment

k33 ¼ ri ;

2

ð4Þ

ð4Þ

ð4Þ

i ¼ 1; 2; 3; k11 ¼ t4x cos2 a3 þ t4y sin a3 ; k12 ¼ k21 ð4Þ

The present study has been partially sponsored by CIUNSA PROJECT No. 1899.

¼ t4x cos a3 sin a3 þ t4y sin a3 cos a3 ; k22 2

ð4Þ

¼ t4x sin a3 þ t 4y cos2 a3 ; k33 ¼ r4 : The application of (16) to (11) leads to

Appendix A Rigidity coefficients of the functional defined by (10) and algebraic procedure to obtain the expression (21). As a consequence of the existence of the angles of inclination ai, i = 1, 2, 3, the rigidity coefficients tix and tiy generate the following coefficients included in Eq. (10): 2

ðiÞ

k11 ¼ t ix cos2 ai þ t iy sin ai ; ðiÞ

ðiÞ

k12 ¼ k21 ¼ t ix cos ai sin ai þ tiy sin ai cos ai ; ðiÞ

2

k22 ¼ t ix sin ai þ t iy cos2 ai ;

3 Z X

tb

Z li 

~i @ui @u @wi ðxi ;tÞ ðxi ;tÞ þ qi Ai ðxi ;tÞ @t @t @t i¼1 " Z Z 3 tb li X ~i ~i @w @ui @u @ 2 wi Ei Ai ðxi ;tÞ ðxi ;tÞ þ Ei Ii 2 ðxi ;tÞ ðxi ;tÞ dxi dt  @t @x @x @xi i i t 0 a i¼1 # 3 Z tb h 2 X ~i @ w ðiÞ ~ i ð0;tÞ þ kðiÞ ~ ðxi ;tÞ dxi dt  k11 ui ð0;tÞu 22 wi ð0;tÞwi ð0;tÞ @x2i i¼1 t a ~i @w ðiÞ @wi ðiÞ ~ i ð0;tÞ þ kðiÞ ~ ð0;tÞ ð0;tÞ þ k12 ui ð0;tÞw þ k33 12 wi ð0;tÞui ð0;tÞ dt @xi @xi Z tb  ð4Þ ð4Þ @w3 ~ 3 ðl3 ;tÞ þ kð4Þ ~ k11 u3 ðl3 ;tÞu ðl3 ;tÞ  22 w3 ðl3 ;tÞw3 ðl3 ;tÞ þ k33 @x3 ta ~3 @w ð4Þ ~ n ðl3 ;tÞ þ kð4Þ ~ ðl3 ;tÞ þ k12 u3 ðl3 ;tÞw 12 w3 ðl3 ;tÞu3 ðl3 ;tÞ dt: @x3

~Þ ¼ dFðv ; v

ta

0

qi Ai

334

R.O. Grossi, C.M. Albarracín / Applied Acoustics 74 (2013) 325–334

Using integration by parts we obtain 3 Z X

tb

Z

li

"

2

2

2

@ ui @ w @ u ~ i þ qi Ai 2 i w ~ i  Ei Ai 2i u ~i u @xi @t 2 @t # " 3 Z tb X ~ i ð0;tÞ @ 4 wi @ 2 wi ð0;tÞ @ w ~ i dxi dt þ þ Ei I i 4 w Ei I i 2 @xi @xi @x t a i i¼1

~Þ ¼  dFðv ; v

i¼1

 Ei I i

ta

0

qi Ai

~ i ðli ;tÞ @ 2 wi ðli ;tÞ @ w @ 3 wi ðli ;tÞ ~ i ðli ;tÞ þ Ei I i w 2 @xi @xi @x3i

# @ 3 wi ð0;tÞ @ui @ui ~ ~ ~ ð0;tÞ þ E A ð0;tÞ u ð0;tÞ  E A ðl ;tÞ u ðl ;tÞ dt w i i i i i i i i i @xi @xi @x3i 3 Z tb  X ðiÞ ðiÞ @wi ~ i ð0;tÞ þ kðiÞ ~  k11 ui ð0;tÞu ð0;tÞ 22 wi ð0;tÞwi ð0;tÞ þ k33 @xi t a i¼1 ~i @w ðiÞ ~ i ð0;tÞ þ kðiÞ ~ ð0;tÞ þ k12 ui ð0;tÞw 12 wi ð0;tÞui ð0;tÞ dt @xi Z tb  ð4Þ ð4Þ @w3 ~ 3 ðl3 ;tÞ þ kð4Þ ~ k11 u3 ðl3 ;tÞu ðl3 ;tÞ  22 w3 ðl3 ;tÞw3 ðl3 ;tÞ þ k33 @x3 ta ~3 @w ð4Þ ~ 3 ðl3 ;tÞ þ kð4Þ ~ ðl3 ;tÞ þ k12 u3 ðl3 ;tÞw 12 w3 ðl3 ;tÞu3 ðl3 ;tÞ dt: @x3 Ei Ii

Because the admissible directions satisfy the compatibility Eqs. (37), (38), (40) and (41) it is possible to collect terms adequately and the use of the differential operators (18)–(20) evaluated at (ai, t) with the following coefficients: bi ¼ Ei Ii ; c1 ¼ 0; c2 ¼ b1 ; c3 ¼ b2 ; b4 ¼ b3 ; c4 ¼ 0; di ¼ Ei Ai ; i ¼ 1;2;3; d4 ¼ d3 ; e1 ¼ 0; e2 ¼ sinða2  a1 ÞE1 I1 ; e3 ¼ sinða2  a3 ÞE2 I2 ; f 1 ¼ 0; f 2 ¼ G1 E1 A1 ; f3 ¼ G2 E2 A2 ; e4 ¼ d3 ; e4 ¼ 0; f 4 ¼ 0; g 1 ¼ 0; h1 ¼ 0; g 2 ¼ cosða2  a1 ÞE1 I1 ; h2 ¼ G3 E1 A1 ; g 3 ¼ cosða2  a3 ÞE2 I2 ; h3 ¼ G4 E2 A2 ; g 4 ¼ 0; h4 ¼ 0:

In all cases, is j = i if i = 1, 2, 3 and j = i  1 if i = 4. sin a2  cos a1 sinða2  a1 Þ cos a3  sin a2 sinða2  a3 Þ ; G2 ¼ ; sin a1 cos a2 cos a2  cos a1 cosða2  a1 Þ sin a2 cosða2  a3 Þ  sin a3 G3 ¼ ; G4 ¼ : sin a1 cos a2 G1 ¼

References [1] Troutman JL. Variational calculus and optimal control. New York: SpringerVerlag; 1996. [2] Sagan H. Calculus of variations. New York: Mc Graw Hill Book Co.; 1969. [3] Weinstock R. Calculus of variations with applications to physics and engineering. New York: Dover; 1974. [4] Gelfand IM, Fomin SV. Calculus of variations. New Jersey: Prentice Hall, Englewood Cliffs; 1963. [5] Dym C, Shames I. Solid mechanics: a variational approach. New York: Mc Graw Hill Book Company; 1973.

[6] Kantorovich L, Krylov V. Approximate methods of higher analysis. New York: Interscience Publishers; 1964. [7] Hildebrand F. Methods of applied mathematics. New Jersey: Prentice Hall; 1965. [8] Elsgoltz L. Calculus of variations. Mass: Addison-Wesley Publishing Co. Reading; 1962. [9] Blevins RD. Formulas for natural frequency and mode shape. New York: Van Nostrand Reinhold; 1979. [10] Karnovsky IA, Lebed OI. Free vibrations of beams and frames. New York: Mc Graw Hill Engineering Reference; 2004. [11] Clough RW, Penzien J. Dynamics of structures. New York: Mc Graw Hill Book Company; 1975. [12] Nowacki W. Dynamics of elastic systems. New York: John Wiley and Sons; 1963. [13] Craig R. Structural dynamics. An introduction to computer methods. New York: John Wiley and Sons; 1997. [14] Timoshenko S, Young DH. Vibration problems in engineering. New Jersey: D. Van Nostrand Company, Inc.; 1956. [15] Chang CH. Mechanics of elastic structures with inclined members. Berlin: Springer-Verlag, The Netherlands; 2005. [16] Oguamanam DC, Hansen JS, Heppler GR. Vibrations of arbitrarily oriented twomember open frames with tip mass. J Sound Vib 1998;209:651–69. [17] Heppler GR, Oguamanam DC, Hansen JS. Vibrations of a two-member open frame. J Sound Vib 2003;263:299–317. [18] Chen WH, Wu CW. A spline wavelets element method for frame structures vibration. Comput Mech 1995;16:11–21. [19] Lee HP, Ng TY. In-plane vibrations of planar frames structures. J Sound Vib 1994;172:420–7. [20] Baychev I. Finite element method for frames with variable characteristics. Mech Res Commun 1996;23:213–20. [21] Chui PPT, Chan SL. Vibration and deflection characteristics of semi-rigid jointed frames. Eng Struct 1997;19:1001–10. [22] Mead DJ. Free vibrations of self-strained assemblies of beams. J Sound Vib 2002;249:101–27. [23] Lin HP, Ro J. Vibration analysis of planar serial-frame structures. J Sound Vib 2003;262:1113–31. [24] Wang X, Wang Y, Zhou Y. Applications of a new differential quadrature element method to free vibrational analysis of beams and frames structures. J Sound Vib 2004;269:1133–41. [25] Yuan S, Kangsheng Y, Williams FW. Second order mode-finding method in dynamic stiffness matrix method. J Sound Vib 2004;269:689–708. [26] Filipich CP, Laura PA. In-plane vibrations of portal frames with end supports elastically restrained against rotation and translation. J Sound Vib 1987;117:467–73. [27] Filipich CP, Valerga de Greco BH, Laura PA. A note on the analysis of symmetric mode of vibrations of portal frames. J Sound Vib 1987;117:198–201. [28] Laura PA, Valerga de Greco BH, Filipich CP. In-plane vibrations of frames carrying concentrated masses. J Sound Vib 1987;117:447–58. [29] Albarracín CM, Grossi RO. Vibrations of elastically restrained frames. J Sound Vib 2005;285:467–76. [30] Grossi RO, Albarracín CM. A variational approach to vibrating frames. J Multibody Dynam 2007;221:247–59. [31] Zeidler E. Applied functional analysis: main principles and their applications, vol. 109. New York: Springer-Verlag; 1995. [32] Reddy JN. Applied functional analysis and variational methods in engineering. New York: Mc Graw-Hill Book Co.; 1986.